• Introducing Dark Mode! Switch by clicking on the lightbulb icon next to Search or by clicking on Default style at the bottom left of the page!

Running UBE Locally vs. Server

tet_hadavas

Active Member
Dear List,

I encountered something strange. I made a copy of a version of R43530 and modified it. It ran the way I wanted on my workstation. So I built an update package and deployed it to the Enterprise Server. When I ran this UBE through the server, a couple key information (e.g. the buyer and phone numbers) were missing. Has anyone seen this before? Any help will be greatly appreciated.

Thanks,
Tet

OneWorld Xe, SP15, Windows 2000, SQL 7.0
 

kastanek

Well Known Member
Hello! Thank you! YES! YES! YES!

3 weeks ago we were the only ones that had reported this issue and now it
seems that more and more are coming out of the wood work. We're having
this issue with LOTS of UBEs and the problem seems to be spreading.

It appears that when submitting various UBE's to the server neither the
Processing options nor the Data Selection information is getting to the
server. We've run DEBUG on both Server and Local runs and on the server
the jobs provides the wrong results and the DEBUG logs shows that the SQL
statement is incorrect (ie. Var = ' ') or that a BSFN Call
(JDB_SetSelectionX) has failed whereas the Local run gives the correct
results and the Local DEBUG log shows the correct SQL statement (ie. Var
= 'data') or that a BSFN Call (JDB_SetSelectionX) has succeeded.

We've done extensive testing in this area and JDE has pulled a lot of
people on this issue but so far we have had no success pinning down the
ultimate cause. Our findings are based on observation and does not seem
to be environemnt specific. Some UBE's do not run in PY but do in DV, JD
and PD. Others do not run in DV or PY but do in JD and PD.

If anyone has this issue I think JDE needs to be made aware of it. There's
power in numbers and the more examples we have the better chance we have of
finding a common thread. Please report your issues to JDE Help Desk.

So far we have identified the following UBE's:

R03B707
R04413
R04570
R09321
R09410
R09420
R09706
R09801
R09813
R10211B
R43510

If we can find a common, BSFN, ER, DLL, etc. then maybe we can find a cure
faster.

Regards,
Gerald







tet_hadavas
<tet.hadavas@suppl To: jdelistml@jdelist.com
ycore.com> cc:
Sent by: Subject: Running UBE Locally vs. Server
owner-jdelistml@jd
elist.com


08/13/2001 04:15
PM
Please respond to
jdelist





Dear List,

I encountered something strange. I made a copy of a version of R43530 and
modified it. It ran the way I wanted on my workstation. So I built an
update package and deployed it to the Enterprise Server. When I ran this
UBE through the server, a couple key information (e.g. the buyer and phone
numbers) were missing. Has anyone seen this before? Any help will be
greatly appreciated.

Thanks,
Tet

OneWorld Xe, SP15, Windows 2000, SQL 7.0




--------------------------
 
We applied SP16, SP16_003, SP16_007, Update-2 and other related ESUs last week. Since then, our PO printing program does not update the next status code to 380 though the processing option "Status Update" is set to blank (Update automatically) when run it on the server. We we run it locally it works perfectly.

XE, SQL2000, SP16, UPD2, Win2000
 
We have a similar issue that happened only in our Prod environment, but now
affects CRP as well.....Another symptom is that the page number and time
don't show up on the reports, as well as some data missing that should be
there. Does this occur for anyone else?

Thanks,

B7332
SP 15.1
Unix 11.0 HP flavor
Oracle
 

Peekaboo

Member
We're also having the same problems that you have described. All UBE's that are submitted through PD to the server complete successfully but either show no data selected or in the case of the R014021 One Line per Address data only appears in only 3 of the 9 or so columns. The issue first appeared when we copied the PY pathcode into the PD pathcode to "refresh" production which until that time had not been modified. We did the refresh at Oracle level rather than using the R98403. This issue has been logged with JDE but we are going live in approximately 2 weeks and need a resolution ASAP.

In light of what has been posted here, we have to consider whether to advise our client to delay "Go-Live" because we don't know that if we use another environment instead of PD that this "problem" won't spread.

If anyone could advise me of the JDE call reference under which they logged this issue I would be obliged as I can't find any calls on the KG about it.

Thanks.

OneWorld Xe SP15.1, Update 2, HP-UX 11.0, Oracle 8.1.7, Citrix, Windows 2000
 

kastanek

Well Known Member
Hi Paul,

We were having this issue with lots of UBE's and everything we did pointed
to an SQLPKG issue but just deleting them was not enough. One of my
collegues found a document on the KG that told us to add a few steps to the
deletion process. I guess the AS/400 is too smart for it's own good and
keeps things in cache even if you want them deleted.

Here's what we did:

The document said to:

- End the JDE Services (remove JDE locks on objects)
- CLRIPC
- ENDSBS QSERVER (remove locks from user jobs but we just put the system
in a restricted state to remove all possible issues)
- Deleted ALL OW related SQLPKG's (Do NOT delete any SQLPKG's that begin
with Q* or _* as those are O/S related)
- IPL the machine
- Restart JDE srevices (if not set to start automatically on IPL)

This procedure cleared up our issues of jobs running correctly locally but
not when submitted to the server. It was a long raod but eventually the
issue was taken care of.

I hope this helps.

Regards,

Gerald.






Peekaboo
<paul_leopard@Hot To: jdelistml@jdelist.com
mail.com> cc:
Sent by: Subject: Re: Running UBE Locally vs. Server
owner-jdelistml@j
delist.com


08/22/2001 06:46
AM
Please respond to
jdelist





We're also having the same problems that you have described. All UBE's that
are submitted through PD to the server complete successfully but either
show no data selected or in the case of the R014021 One Line per Address
data only appears in only 3 of the 9 or so columns. The issue first
appeared when we copied the PY pathcode into the PD pathcode to "refresh"
production which until that time had not been modified. We did the refresh
at Oracle level rather than using the R98403. This issue has been logged
with JDE but we are going live in approximately 2 weeks and need a
resolution ASAP.

In light of what has been posted here, we have to consider whether to
advise our client to delay "Go-Live" because we don't know that if we use
another environment instead of PD that this "problem" won't spread.

If anyone could advise me of the JDE call reference under which they logged
this issue I would be obliged as I can't find any calls on the KG about it.

Thanks.

OneWorld Xe SP15.1, Update 2, HP-UX 11.0, Oracle 8.1.7, Citrix, Windows
2000
--------------------------
Visit the forum to view this thread at:
http://198.144.193.139/cgi-bin/wwwthreads/showflat.pl?Cat
=&Board=OW&Number=19425
 

NagapA

Member
We have a similar problem with running UBE's locally and on the server. We use the R46471, which calls R46472 (Print Move slips) and in turn calls R46473 (Print Move Tags).

If we run the process locally, the client workstation will process the R46471 46472 and 46473 and create the PDFs. However, when we run these processes on the server, R46472 will generate the PDF and will not generate a PDF for R46473. We have tried everything, including rebuilding another Full Package and deploying it, unfortunately it has not done anything for us.

What we do now is generate the Move Slips and Tags locally and than run the application to reprint the Move Slips and Tags on the server and that somehow generates the PDFs we need. Any one else facing similar issues? Please help.

We need the PDF's generated on the Enterprise server becasue we use a third party application caleed JetForms which in truns takes the PDF and prints the move tags for us on a Zebra150 printer.
 

NoSleep

Member
I have the same issue with my environments and its driving me mad.

I've done some testing and have noticed one thing in particular, maybe a clue:

I built a custom UBE on top of the same views as one of the UBE's that are not working and built it on the server. Low and behold, the ube runs fine and retrieves all my data correctly.

So what does this tell me? I'm thinking that it has to be a dll since I didn't use any bsfn's in the ube design. What do y'all think?

NoSleep

Xe, SP13, Update 3, Win2k, SQL2000
 

husni

Well Known Member
We applied SP17 and Update 3 recently and built a fresh server and client
package. We found the pain in the worst part of
the body is over now. I am not sure if the SP17 or Update 3 resolved the
problem.

Husni
Xe, SP17, Update 3
Win2k, SQL2000
 

NoSleep

Member
Guys, I've made some progress on this issue (at least I think I did).

On a chance to see if maybe my specs on the enterprise server was corrupt, I replaced my bin32 directory for the environment that was giving me the ube problem with a bin32 directory from another environment that was not giving me the ube problem. Low and behold, the ube started working.

When I compared the object files of the package builds on the enterprise server (the problem package build that I'm using and an older build), I noticed that some of my BSFN's were not being compiled. I also saw that there was a text file in each of the dll folders (e.g. CAEC) for the package in question and saw that the linking was not being done correctly.

I then compared the package builds being done on my deployment server vs. my enterprise server (two separate boxes) and noticed that again the pacakge on my enterprise server was missing some bsfn's in its compile.

So, what does this tell me. That there is something definitely wrong with the bsfn compile on my enterprise server. its not the specs, its the C stuff.

I'm looking at all my environment settings and ini settings and am going to do a full build again tonight.

I hope this info helps you guys with these similar issues.

NoSleep

OneWorld Xe, Update 3, SP13, Win2k, SQL2000
 

NoSleep

Member
Alright, new info on this issue. Folks, just tell me if y'all have heard of this already.

I looked at my logs again closely, the logs on the enterprise server in the obj folder for each of my dlls. I notice that some of the references to jdekrnl.lib are not being made. It looks like jdekrnl.lib actually has the code for some of the JDE APIs and for some reason, has not been calling that code correctly when doing the compilation. My thinking is that the system files on the enterprise server are corrupt, including the libraries, hence my build problems and the reason why some of my object code is not getting compiled. I talked with another CNC buddy and he said that he has seen this issue once on NT and another on the AS/400(he's actually working on the 400 one right now).

This also explains why everytime I do full server package that I get the same issue with the UBE's not running on the server correctly (R014021 is only showing data for 3 of 10 columns).

So my plan is:

1. replace the system folders from backup (after I call JDE and get some questions answered re: the registry)
2. execute full package.

I think this will resolve the UBE issue. I'll let yall know.

NoSleep

OneWorld Xe, Update 3, SP13, Win2k, SQL2000
 

adamd

Active Member
Anybody make any progress on this slower on the server issue ?

Thanks

OneWorld XE, SP 14.2
HPUX 11, Oracle 8.0.6
 
Top